On Wed, Mar 21, 2012 at 6:16 AM, Noah Slater <[email protected]> wrote: > I noticed this, along with a few other changes, made to the 1.2.x branch, > that do not seem related to the release blockers. >
Its related to COUCHDB-1445. Not sure about the other ones. > Randal, or anyone else, could you explain what they are for? > You'll have to be more specific for me to comment on anything else. > On Wed, Mar 21, 2012 at 9:17 AM, Jason Smith <[email protected]> wrote: > >> Yeah, it worked out, thanks. The frustrating problem with incorrect >> `receive` patterns is >> >> On Wed, Mar 21, 2012 at 8:49 AM, Randall Leeds <[email protected]> >> wrote: >> > On Wed, Mar 21, 2012 at 01:34, Randall Leeds <[email protected] >> >wrote: >> > >> >> Thanks for cleaning up after me. Sloppy, Randall. Bad, Randall. >> >> >> > >> > Well, I was mostly right... >> > >> > >> >> >>> In particular, the error is with the change to couch_log.erl. >> >>> >> >> >>> >> Whoops. I meant couch_file.erl. >> >>> >> >> >>> >> Randall, please see the attached diff, or see >> >>> >> http://friendpaste.com/2Hwjn9uj1oRM8pyRqAdbxc >> >>> >> >> >>> >> The pattern {Ref, Pid, {error, Reason}} does not match because the >> >>> >> message is {Ref, Pid, Reason} >> >>> >> >> >> >> >> > Only in one case, where we detect that the file exists. >> > I've changed it to return {error, eexist} in this case, which matches the >> > errors that come back from the file module in all other places and will >> > format the correct message. >> > >> > Fixed on 1.1.x and 1.2.x. >> >> >> >> -- >> Iris Couch >>
